Points of View Status of inclusion of EQ-5D in clinical trial applications in each country

Printable PDF

Yuki Miura, Senior Researcher, Pharmaceutical and Industrial Policy Research Institute

Executive Summary

The percentage of EQ-5D incorporated in clinical trials was high in the U.K., Japan, Germany, France, and the U.S., in that order. The low inclusion rate of the EQ-5D in the U.S. was attributed to the low inclusion of the EQ-5D in phase 2, where the total number of clinical trial applications was high.

The simultaneous inclusion of the EQ-5D and other indicators was about 10% for both labor productivity-related indicators and caregiving-related indicators when compared before Phase 3 and in Phase 4. Compared to the total number of clinical trial applications, a higher percentage of labor productivity-related indicators were concurrently included in the EQ-5D compared to caregiving-related indicators.

In terms of whether evidence obtained after market launch could be used for evaluation, further analysis of Phase 4 showed that the majority of each relevant indicator was measured by Secondary Outcome Measures. It is thought that the hurdle for obtaining these outcomes itself may not be high by incorporating a questionnaire without the main objective of measuring these measures.

1. Introduction

The Pharmaceutical Manufacturers Association of Japan's (hereafter "PMAJ") Pharmaceutical Manufacturers Association of Japan (hereafter "PMAJ") Policy Recommendation 2023 argues that a system is needed to reflect post-launch value assessment in drug prices, such as reassessing values based on evidence obtained after the product is launched and changes in positioning in guidelines, etc. ⁽¹⁾. In the past, there have been products whose true clinical usefulness was verified after their ^launch2), but these products were evaluated for reduction in the risk of cardiovascular events, etc. separately from clinical trials for the purpose of marketing approval, and received an addition to the drug price at the time of drug price revision.

In light of recent healthcare resource allocation perspectives, it would be useful from the perspective of healthcare resource allocation to present evidence from a healthcare economic perspective in addition to conventional clinical usefulness and to evaluate such evidence. However, under the current drug pricing system, the only system that evaluates the results of health economic analysis is the cost-effectiveness evaluation system. The cost-effectiveness evaluation system is complementary to the NHI drug price system and is limited in its implementation, especially for drugs that have a large financial impact. Uncertainty is a problem in the cost-effectiveness evaluation system because the calculations are made with many assumptions and estimates that arise during the analysis, causing variation in the results. We believe that it is important to clarify the reality of the simultaneous inclusion of outcomes, since obtaining these data for the same population will reduce the uncertainty that arises during the analysis.

In addition, the ISPOR (International Society for Pharmacoeconomics and Outcomes Research) Value flower lists quality-adjusted life These two elements are one of the various values of pharmaceutical products. 3 ⁾ In Japan, it has been reported that Lecanemab has the potential to improve health outcomes and quality of life (QOL) for patients with early-stage Alzheimer's disease and their caregivers, as well as to reduce their financial ^burden4). The study was conducted using a simulation model4) to examine the impact of the disease from both a payer perspective, focusing on direct care costs (such as outpatient and inpatient services, long-term care and home health care services, party pharmaceutical costs, and other intervention costs), and a social perspective (social costs including informal care costs due to family caregivers in addition to direct care costs). ^The results of these analyses are expected to be evaluated at the time the NHI drug price is listed on the market.

In this paper, we focus on health-related quality of life (EQ-5D in this paper), which is used in health economic analysis, and labor productivity and reduced burden of caregiving, two components of ISPOR's valueflower for which outcome measures exist in a previous ^{study5) by} Nakano et al. The purpose of the survey was to determine the extent to which these outcome measures are incorporated into clinical trial applications (including simultaneous incorporation with the EQ-5D) and the differences in the incorporation of outcome measures before and after the launch of the product.

2-1. survey target

The third edition of the Central Social Insurance Medical Council's Guidelines for Analyzing Cost-Effectiveness Evaluation recommends the EQ-5D-5L as the first ^choice6) when collecting new QOL values in Japan for the purpose of cost-effectiveness analysis. Therefore, this paper investigates the actual status of clinical trials that incorporate the EQ-5D as an outcome measure.

The EQ-5D can be used in clinical trials, observational studies, population health surveys, routine outcome measures, and many types of studies where general measures of health status are useful, and EQ-5D index values can be used to estimate quality-adjusted life years (QALY The position of the EQ-5D in other countries is that it is the recommended scale for assessing health-related quality of life in adults by NICE in the UK, accounting for 49% of all measures used in ^{assessments8)} (other ^HUI9): 13%, SF-6D: 1%, disease-specific quality of life rating scale 19 ). Although disease-specific quality of life rating scales are also widely used, they were not included in this study because the main purpose of this study was not to conduct a disease-specific analysis.

In addition, not all patient-reported outcomes and QOL measures can be used to calculate QOL values used in cost-effectiveness analyses, but only those measured by preference-based scales developed to calculate ^QALYs6) as shown in Figure 1. Although the HUI and SF-6D can be used to calculate QALYs, the HUI and SF-6D are not included in this study because the EQ-5D, which is widely and commonly used in cost-effectiveness analysis, is used in this study. Labor productivity and reduced burden of caregiving are described in the next section of the survey methodology.

2-2. survey method

To investigate the actual status of inclusion of EQ-5D clinical trials in Japan, the U.S., the U.K., Germany, and France at the time of application, data from ClinicalTrials.gov, a U.S. clinical trials database that provides a comprehensive overview of the status of clinical trials in each country, was used to compile the data. The search conditions for data extraction were as follows.

ClinicalTrials.gov search conditions

The data obtained with the above search conditions were cleaned as follows.

In the "Outcome Measures (hereafter referred to as "Outcome")," "EQ-5D," "EQ5D," "European Quality of or "EuroQOL" were considered as trials that incorporated "EQ-5D" and were included in the total number of trials. In addition, those with no data in "First Posted" at the time of analysis and those with an application period other than 2010.01.01 to 2022.12.31 were excluded from the analysis, and only those that matched the period in the search criteria were included. In the "Intervention" category, those that contained either the word "Drug" or "Biologic," and in the "Study Phase" category, those that contained the word "Phase 1" alone were included. In "Location," the words in parentheses were included from the United States (United States), the United Kingdom (United Kingdom), Germany (Germany), France (France), and Japan (Japan). As of September 20, 2023, there were 94,901 trials that met the search criteria, and 1,420 trials were included in the analysis as a result of data cleaning.

Among the "Outcome," the outcome indicators of labor productivity (hereinafter referred to as "labor productivity-related indicators") were researched and examined by Nakano et al. ¹⁰⁾ "WPAI" and "productivity" were used as search terms, and "family caregiver indicators" were used as search terms. The family caregiving indicators (hereafter referred to as "caregiver-related indicators") were also investigated and examined by Nakano et al. ¹¹⁾ "Caregiver," "Carer," and "Family" were used as search terms, 11) We also investigated the relationship with the EQ-5D using "Caregiver," "Carer," and "Family" as search terms.

Results

Figure 2 shows the extent to which EQ-5D was included in clinical trials in each of the countries surveyed in this study.

Looking at the cumulative number of trials incorporating the EQ-5D over the study period 2010-2022, the United States had 875, the United Kingdom 668, Germany 626, France 625, and Japan 328 (Figure 2). The United States had the largest number of cases with 875, while Germany, France, and the United Kingdom had between 625 and 668, with no significant differences among the European countries. Japan had the fewest among the five countries with 328 cases.

The percentage of EQ-5D inclusion per country was calculated for the data after the cleaning was conducted and for the other data cleaning excluding EQ-5D under the same conditions. The plot in Figure 2 shows that the U.S. had the highest number of EQ-5D included studies (875), but the lowest percentage (2.5%). The percentage was 10.1% in the United Kingdom, 8.4% in Germany, 7.9% in France, and 8.7% in Japan, with the United Kingdom, Japan, Germany, France, and the United States having the highest percentages in that order.

In order to confirm the reason for the low percentage of EQ-5D inclusion studies in the U.S., the percentage of EQ-5D inclusion studies by phase and the number of studies filed were reviewed (Figure 3). The bar graph shows the number of U.S. studies that did not incorporate EQ-5D during the period, and the plot shows the percentage of studies that incorporated EQ-5D in the U.S. Phase 2 had a lower percentage of studies (1.6%) than the other phases, partly due to the large number of applications. On the other hand, Phase 3 had a higher percentage of trials (6.3%) than the other phases. In this study, these were combined to secure the sample, and Figure 4 confirms whether Phase 3 alone and the others show different trends.

From Figure 4, we confirmed whether there were differences by Phase 3 in each country. The difference between the percentage of EQ-5D inclusion trials in Phase 3 and the same percentage in Figure 2 shows that only in the U.S. there is a large gap, which is attributed to the low percentage in Phase 2, which has a larger population.

Due to sample issues in the subsequent analysis, the five countries in Figure 2 were combined for the analysis. Next, we analyzed whether the clinical trials under investigation fell into the Drug or Biologic category.

When Drug and Biologic were considered as two different modalities, Drug was identified as so-called small molecule drugs and Biologic as high molecular weight drugs. Of these, Drug accounted for 1,257 studies and Biologic for 170 studies, indicating that Drug accounted for the majority of studies; when looking at the composition of the data without the inclusion of EQ-5D (other data cleaning excluding EQ-5D was done under similar conditions), Drug accounted for 87%, Biologic was 13% (Supplementary Data 1). This is a very small difference compared to 88% for Drug and 12% for Biologic with the inclusion of EQ-5D, indicating that there is no difference in EQ-5D acquisition trends between Drug and Biologic. Note that there are cases in which both Drug and Biologic are applicable in the same study, and these overlaps do exist.

Next, in order to investigate the extent to which the EQ-5D, labor productivity-related indicators, and care-related indicators are incorporated in clinical trials before and after a product is launched, we conducted an analysis by phase of clinical trials in the five countries in which each indicator was incorporated (other data cleaning conditions were similar except for the EQ-5D). The analysis was conducted in the following way.

The "EQ-5D" was divided into two groups: Phase 3 or earlier (excluding Phase 1 alone) or Phase 4 of the clinical trial based on the data before and after the EQ-5D cleaning (other data cleaning excluding the EQ-5D was conducted under similar conditions), Comparing Phase 3 and Phase 4 data, the absolute numbers of all indicators were higher in Phase 3 and earlier. The majority of the "EQ-5D" cases were in Phase 3 or earlier, with 1,272 cases in Phase 3 or earlier and 148 cases in Phase 4 (the percentage of EQ-5D cases in Phase 3 or earlier: 89.6%), EQ-5D had the highest number of incorporations among the three indicators in both Phase 3 and Phase 4. The number of studies for "labor productivity-related indicators" was 170 before Phase 3 and 66 in Phase 4, which were fewer than for the other indicators. The inclusion of "care-related indicators" was 945 before Phase 3 and 121 in Phase 4.

Next, we analyzed how many caregiving-related and labor productivity-related indicators were simultaneously included in trials that incorporated the EQ-5D.

Figure 7 shows the number of trials in which "labor productivity-related indicators" and "caregiving-related indicators" were simultaneously included in the trials in which the EQ-5D was included in Figure 3, and the percentage of each Before Phase 3, 82 (6.4%) were for labor productivity-related indicators and 129 (10.1%) were for caregiving-related indicators; in Phase 4, 15 (10.1%) were for labor productivity-related indicators and 13 (8.8%) were for caregiving-related indicators. Although statistical analysis was not performed due to the small sample size, prior to Phase 3, the inclusion rate of caregiving-related indicators (10.1%) tended to be higher than the inclusion rate of labor productivity-related indicators (6.4%); in Phase 4, the inclusion rate of caregiving-related indicators (10.1%) tended to be higher than the inclusion rate of labor productivity-related indicators (10.1%). Comparing the percentage of each indicator in Phase 3 and 4, there was no significant difference in the percentage in each phase, but in Phase 3 and before, the percentage of caregiving-related indicators was higher than that of labor productivity-related indicators. However, there was a difference in that the ratio of caregiving-related indicators was higher than that of labor productivity-related indicators before Phase 3, and in Phase 4, the ratio of caregiving-related indicators tended to be slightly lower than that of labor productivity-related indicators.

Figure 8 shows the percentage of the total number of labor productivity-related indicators and caregiving-related indicators in Figure 6 that were simultaneously incorporated with the EQ-5D. In Phase 4, 22.7% (15/66) of the cases were measured at the same time, while in Phase 4, the percentage was lower. In the case of "care-related indicators," the percentage of simultaneous inclusion was 13.7% (129/945 cases) before Phase 3 and 10.7% (13/121 cases) in Phase 4, both of which were lower than that of labor productivity-related indicators (Figure 8).

From the viewpoint of post-marketing evidence, we focused on Phase 4 and investigated which endpoints in the clinical trials were classified as those that were concurrently included in the EQ-5D, to see if this could be an issue for inclusion.

In Figure 9, we checked whether each relevant measure included in Phase 4 was classified as a Primary Outcome Measure or a Secondary Outcome Measure. Those that did not fall under Primary Outcome Measures or Secondary Outcome Measures were excluded. Figure 9 shows that the majority of respondents placed the "EQ-5D," "labor productivity-related indicators," and "caregiving-related indicators" in the Secondary Outcome Measures. The majority of the measures are placed in the Secondary Outcome Measures, indicating that they are included as a secondary outcome in additional clinical trials.

Figure 10 analyzes the number of included trials for each relevant measure by modality (Drug, Biologic) in Phase 4. As in Figure 5, the majority of trials were for Drug rather than Biologic, and there was no difference in this trend.

From the viewpoint of evidence obtained after the product was launched, we analyzed the results of the outcome measures included in Phase 4, "labor productivity-related indicators" and "nursing care-related indicators," in Figure 8, to see which diseases were targeted. In Figure 10, which covers Phase 4, we included those cases in which either the labor productivity-related indicator or the caregiving-related indicator was included at the same time as the EQ-5D. Those with no disease information were considered as not applicable.

Table 1 shows that there were 4 cases of multiple sclerosis, 3 cases of psoriasis, 2 cases each of rheumatoid arthritis, rosacea, and migraine, and 1 case each of "labor productivity-related indicators" or "nursing care-related indicators" thereafter. The most common type of disease was multiple sclerosis.

Multiple sclerosis, the most common disease, is a chronic inflammatory demyelinating disease of the central nervous system and is designated as an intractable disease in Japan. The main symptoms include visual impairment, diplopia, cerebellar ataxia, paralysis of the limbs (monoplegia, paraplegia, hemiplegia), sensory disturbance, cysto-rectal disturbance, gait disturbance, and painful tonic spasm, depending on the site of ^lesion12). There is also a FAMS (Functional Assessment of MultipleSclerosis) ¹³ that measures health-related quality of life in multiple sclerosis, but it was not simultaneously incorporated in the present case.

Conclusion and Discussion

In this paper, we focus on health-related quality of life (EQ-5D in this paper), which is used in health economic analysis, and labor productivity and reduced burden of caregiving, two components of ISPOR's Valueflower for which outcome measures exist in previous studies by Nakano et al. The purpose of the survey was to determine the extent to which these outcome measures were incorporated at the time of clinical trial application, and the differences in the incorporation of these outcome measures before and after the launch of the product.

The results showed that the United States had the largest number of trials incorporating the EQ-5D, followed by the three European countries of the United Kingdom, Germany, and France, and then Japan. The United States had a large number of Phase 2 clinical trial applications (15,528), and the inclusion rate of EQ-5D was low at 1.6%. In the other countries, we inferred that there may have been many trials with simultaneous inclusion due to international joint trials, etc. We considered that there was an overlap between countries, resulting in a smaller difference in the inclusion rate in countries other than the U.S. In addition, we speculated that the inclusion rate was higher in the U.K. than in Japan, Germany, and France because the U.K. uses the results of HTA (Health Technology Assessment) in making decisions on reimbursement, if only for EQ-5D. and Phase 4 in countries where HTA results are used in reimbursement decisions and those where they are not, but due to the large number of duplicate trials, we were not able to examine these factors. In addition, it should be noted that ClinicalTrials.gov, which was the subject of the survey, is a U.S. database and does not include information on all clinical trials conducted or underway in other countries, including Japan. ¹⁴⁾ Therefore, clinical trials conducted outside the U.S. may be underestimated. Therefore, clinical trials conducted outside of the U.S. may be underestimated14) . Although not covered in this survey, when evaluating Japanese clinical trials themselves, there is room for consideration of information on clinical trials and clinical research in Japan in comparison with information on the Clinical Research Information Portal ^{Site15) ,} UMIN (University Hospital Medical Information Network) ¹⁶⁾, JAPIC (Japan Pharmaceutical Information Center) ¹⁷⁾, etc. There is room to consider the comparison of information from the following sources.

Regarding the modalities in the big picture, our hypothesis before the survey was that Biologic incorporated more of these outcomes into clinical trials, but no difference was found. Although the EQ-5D, which calculates Utility for cost-effectiveness analysis from a health economic perspective, was included in the survey, not all QOL values used in the analysis had to be incorporated into clinical trials; they could be obtained from prior papers or databases18) , and not all drugs were It is possible that this is a limitation because it was not necessary to incorporate them into new clinical trials. In addition, disease-specific QOL evaluation scales were not surveyed in this study, and considering the cases in which QOL values were not measured by the EQ-5D, we believe it is necessary to include trends in QOL evaluation scales themselves.

In Phase 4, 15 cases (10.1%) were for labor productivity and 13 cases (8.8%) were for caregiving-related indicators, both around 10%. In Phase 4, the labor productivity-related indicators were 15 (10.1%) and caregiving-related indicators were 13 (8.8%), both around 10%.

Looking at the value obtained by dividing the number of studies in which the EQ-5D was simultaneously incorporated by the number of studies in which the EQ-5D was not incorporated (Figure 8), labor productivity-related indicators were incorporated at a relatively higher rate than care-related indicators, while care-related indicators tended to be incorporated at a lower rate of around 10%. Although it is not possible to determine whether this is due to the use of disease-specific quality of life evaluation scales other than the EQ-5D in this survey, or whether it is due to the subject of the data, it is possible to reduce uncertainty in the analysis if these data are obtained in the same population from the perspective of health economic analysis, and therefore, the simultaneous inclusion of these indicators in the EQ-5D was not included in this survey. We believe that further analysis of factors that are not incorporated (not measured) is necessary.

In terms of whether or not evidence acquired after market launch can be used for evaluation, when we proceeded with the analysis for Phase 4, the majority of each relevant indicator was measured by Secondary Outcome Measures. Even if the main objective is not to measure these indicators, the hurdle for obtaining these outcomes itself may not be high by incorporating a questionnaire. Although we were not able to conduct enough analysis in this survey to examine whether these simultaneous measures are being sufficiently measured or whether there is room for improvement, some of the care-related indicators and labor productivity indicators were incorporated, while others were not, and further analysis of the factors that prevented their inclusion was considered necessary. On the other hand, it must be kept in mind that there are some diseases and therapeutic agents that are not amenable to measurement and improvement of QOL and each indicator.

Analysis of the diseases for which each indicator was included in Phase 4 showed that many of them, such as multiple sclerosis, psoriasis, rheumatoid arthritis, and migraine, are diseases in which patients themselves easily identify functional impairment and subjective symptoms. These diseases may be relatively easy to incorporate because they have outcome measures or because they tend to show differences in the quality of life values obtained. If it becomes possible to simultaneously obtain data on multiple outcome measures for the same population, it has the potential to reduce uncertainty in health economic analysis, and more reliable analysis can be conducted. Furthermore, the accumulation of such cases may lead to a higher resolution discussion on the pros and cons of evaluation.

On the other hand, given the fact that there are some diseases and therapeutic agents that are not amenable to measurement, we believe that further discussion is needed on how to include them in future evaluations, including from the perspective of medical resource allocation.

Return to News List

TOP